%pip install -r requirements.txt

Requirement already satisfied: scipy in c:\users\stast\anaconda3\envs\venv\lib\site-packages (from -r requirements.txt (line 1)) (1.7.2)
Requirement already satisfied: wavfile in c:\users\stast\anaconda3\envs\venv\lib\site-packages (from -r requirements.txt (line 2)) (4.4.0)
Requirement already satisfied: rir-generator in c:\users\stast\anaconda3\envs\venv\lib\site-packages (from -r requirements.txt (line 3)) (0.2.0)
Collecting bitstring
  Downloading bitstring-4.0.1-py3-none-any.whl (45 kB)
Requirement already satisfied: cffi>=1.1.0 in c:\users\stast\anaconda3\envs\venv\lib\site-packages (from rir-generator->-r requirements.txt (line 3)) (1.15.0)
Requirement already satisfied: numpy>=1.6 in c:\users\stast\anaconda3\envs\venv\lib\site-packages (from rir-generator->-r requirements.txt (line 3)) (1.21.2)
Requirement already satisfied: pycparser in c:\users\stast\anaconda3\envs\venv\lib\site-packages (from cffi>=1.1.0->rir-generator->-r requirements.txt (line 3)) (2.21)
Installing collected packages: bitstring
Successfully installed bitstring-4.0.1
Note: you may need to restart the kernel to use updated packages.

import numpy as np
import matplotlib.pyplot as plt
import scipy
import scipy.fft

import IPython

from oscillator import SineOscillator, SquareOscillator

NOTES = ["C", "C#", "D", "D#", "E", "F", "F#", "G", "G#", "A", "A#", "B"]
OCTAVES = ["sub-contra", "contra", "great", "small", "1-line", "2-line", "3-line", "4-line", "5-line"]
MAJOR_SCALE_PATTERN = [2, 2, 1, 2, 2, 2, 1]
MINOR_SCALE_PATTERN = [2, 1, 2, 2, 1, 2, 2]

SAMPLE_RATE = 44100

def get_note_number(note, octave):
    if not isinstance(octave, int):
        octave_num = OCTAVES.index(octave)
    else:
        octave_num = octave
    note_num = NOTES.index(note)
    return octave_num * 12 + note_num - 8

def get_note_frequency(note_number):
    return 2 ** ((note_number - 49) / 12) * 440

def get_scale_frequencies(starting_note_number, scale_pattern=MAJOR_SCALE_PATTERN):
    scale_freq = [get_note_frequency(starting_note_number)]
    note_number = starting_note_number
    for step in scale_pattern:
        note_number += step
        scale_freq.append(get_note_frequency(note_number))
    return scale_freq

def get_freq_array_for_scale(scale, oscillator_class):
    NOTE_DURATION_MILLISEC = 1000
    PAUSE_DURATION_MILLISEC = 350
    freq_arr = []
    for note_freq in scale:
        oscillator = oscillator_class(frequency=note_freq, sample_rate=SAMPLE_RATE)
        iterations = SAMPLE_RATE * NOTE_DURATION_MILLISEC // 1000
        iterator = iter(oscillator)
        for i in range(iterations):
            freq_arr.append(next(iterator))
        pause_iterations = SAMPLE_RATE * PAUSE_DURATION_MILLISEC // 1000
        for i in range(pause_iterations):
            freq_arr.append(0)
    freq_arr = np.array(freq_arr)
    return freq_arr

Implemented two types of oscillator: clear sine wave and square wave

IPython.display.Audio(
    data=get_freq_array_for_scale(
        get_scale_frequencies(
            get_note_number("C", "1-line"), 
            MAJOR_SCALE_PATTERN
        ),
        SineOscillator
    ), 
    rate=SAMPLE_RATE, 
    autoplay=True
)

IPython.display.Audio(
    data=get_freq_array_for_scale(
        get_scale_frequencies(
            get_note_number("C", "1-line"), 
            MAJOR_SCALE_PATTERN
        ),
        SquareOscillator
    ), 
    rate=SAMPLE_RATE, 
    autoplay=True
)

IPython.display.Audio(
    data=get_freq_array_for_scale(
        get_scale_frequencies(
            get_note_number("C", "1-line"),
            MINOR_SCALE_PATTERN
        ),
        SineOscillator
    ), 
    rate=SAMPLE_RATE, 
    autoplay=True
)

def fft_plot(audio, sample_rate, max_display_freq=1000):
    n = len(audio)
    T = 1 / sample_rate
    yf = scipy.fft.fft(audio)
    xf = scipy.fft.fftfreq(n, T)
    plt.xlim((0, max_display_freq))
    plt.tight_layout()
    plt.plot(xf, 2 / n * np.abs(yf))

Frequency characteristics of sine wave generated octaves are nice and clean with only base note frequencies having a peak. In square wave we can observe many additional frequencies in the graph which means that square wave can be obtained with addition of multiple sine waves:

fft_plot(get_freq_array_for_scale(get_scale_frequencies(get_note_number("C", "1-line"), MAJOR_SCALE_PATTERN), SineOscillator), SAMPLE_RATE)

fft_plot(get_freq_array_for_scale(get_scale_frequencies(get_note_number("C", "1-line"), MINOR_SCALE_PATTERN), SineOscillator), SAMPLE_RATE)

fft_plot(get_freq_array_for_scale(get_scale_frequencies(get_note_number("C", "1-line"), MAJOR_SCALE_PATTERN), SquareOscillator), SAMPLE_RATE, 10000)